Download A Study on Evaluation of kV-CBCT-image-based Treatment

Survey
yes no Was this document useful for you?
   Thank you for your participation!

* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project

Document related concepts

Positron emission tomography wikipedia , lookup

Radiation therapy wikipedia , lookup

Neutron capture therapy of cancer wikipedia , lookup

Proton therapy wikipedia , lookup

Backscatter X-ray wikipedia , lookup

Nuclear medicine wikipedia , lookup

Radiation burn wikipedia , lookup

Radiosurgery wikipedia , lookup

Fluoroscopy wikipedia , lookup

Sievert wikipedia , lookup

Image-guided radiation therapy wikipedia , lookup

Transcript
Journal of Medical and Biological Engineering, 31(6): 429-435
429
A Study on Evaluation of kV-CBCT-image-based Treatment
Planning using Anthropomorphic Phantom
Padmanaban Sriram*
Nagarajan Vivekanandan
Sukumar Prabakar
Department of Medical Physics, Cancer Institute (WIA), Adyar, Chennai 600036, Tamil Nadu, India
Received 12 Jul 2010; Accepted 8 Nov 2010; doi: 10.5405/jmbe.817
Abstract
Kilovoltage cone-beam computed tomography (kV-CBCT) based on flat-panel technology is primarily used for
patient positioning. It can also be used for re-planning and dosimetric verification in adaptive radiotherapy by validating
the accuracy of CBCT-image-based treatment planning. This study evaluates the accuracy of dose calculation based on
CBCT images using an anthropomorphic phantom. The Hounsfield unit (HU) – electron density calibration curves for
full-fan and half-fan modes of CBCT are obtained using a Catphan® 600 phantom and compared with the conventional
CT curve. The stability of the CBCT calibration curve with time is also studied over a period of 8 weeks. CT and half-fan
CBCT imaging of the anthropomorphic phantom was obtained. The isodose distributions of a single direct field and
wedge fields are compared for CT- and CBCT-image-based dose calculations. Simulated identical targets and organs at
risk are delineated in CT and CBCT images for which conformal radiation therapy (CRT) and intensity-modulated
radiation therapy (IMRT) plans were generated. The percentage dose differences and dose-volume histograms (DVHs) of
CT- and CBCT-based plans are compared. It is found that the maximum difference in HU between half-fan CBCT and CT
is 40 HU. No significant change is observed in the calibration curves over a period of 8 weeks. The isodose distributions
computed based on CBCT and CT for a single direct field and wedge fields agrees to within 1%. For head and neck and
pelvis IMRT plans, the dose calculated using CT agrees with that of CBCT-based calculation to within ± 1.0%,
respectively. However, for the thorax, there is a pronounced discrepancy at the 100% isodose line and the dose difference
is within ± 3%. It can be thus concluded that CBCT images can be used for dose calculation, but it is necessary to validate
the dosimetric data for inhomogeneous tissue regions.
Keywords: Cone-beam computed tomography (CBCT), Dose calculation, Rando phantom
1. Introduction
Recent developments in image-guided radiation therapy
(IGRT) have provided tools for improving patient positioning
and target localization accuracy. Kilovoltage cone-beam
computed tomography (kV-CBCT) integrated into a linear
accelerator was developed to acquire online anatomical and
volumetric images [1]. By matching soft tissues and/or bony
structures in the CBCT to those in planning CT images, patient
positioning can be verified accurately. In addition, to guide the
patient setup, the CBCT data acquired prior to treatment can be
used to recalculate and verify the treatment plan based on the
patient anatomy on the day of treatment [2-6]. This allows the
patient’s treatment plan to be adaptively modified during the
radiation therapy course based on the dose that has already
been delivered. However, this application depends on the
CBCT image quality.
* Corresponding author: Padmanaban Sriram
Tel: +91-9840600601
E-mail: [email protected]
Unlike conventional CT, kV-CBCT covers a much larger
field of view (FOV) in the longitudinal direction. Scatter is thus
a more severe problem in the resultant image. Since the gantry
rotation speed is limited to ~1 rotation per minute by
International Electrotechnical Commission regulations, CBCT is
prone to motion artifacts. Our previous study found that
non-gated CBCT imaging of moving targets can lead to
significant loss of volumetric information but a lesser distortion
of shape [7]. Poor image quality raises serious concerns about
the dosimetric reliability of CBCT-based dose calculations. This
study evaluates the dosimetric accuracy of kV-CBCT-based
dose calculation using an anthropomorphic phantom (Rando
phantom) for the head and neck, thoracic, and pelvic sites.
2. Materials and methods
2.1 Image acquisition
The on-board imager (OBI) integrated into the Varian
Clinac 2100C/D linear accelerator (Varian Medical Systems,
J. Med. Biol. Eng., Vol. 31 No. 6 2011
430
Palo Alto, CA) was used in this work to acquire CBCT images.
The CBCT images were reconstructed using nearly 700
projection images acquired in a single 360-degree rotation of
the gantry. CBCT images can be obtained using the OBI in two
modes: full-fan and half-fan. In the full-fan mode, the beam
central axis passes through the detector center and a full
projection of the scanned patient is acquired for each angle.
This acquisition mode produces a field of view (FOV) of about
25.5 cm in diameter and 17 cm in the axial length. The half-fan
mode is designed to obtain a larger FOV. In this mode, the
detector is shifted laterally to take only half of the projection of
the scanned patient for each angle. This acquisition mode
produces an FOV of about 45 cm in diameter and 15 cm in the
axial length. A bowtie filter was mounted on the X-ray tube to
improve image quality. For imaging, the default parameters,
defined during calibration, were 120 kVp, 80 mA, and 25 ms.
A GE lightspeed CT simulator (GE Medical Systems,
Milwaukee, WI) was used for acquiring CT images with a
2-mm slice thickness. 120 kVp and an automatically adjusted
current were used.
2.2 Calibration of CT and CBCT
To use a CT or CBCT image for dose calculation, the
Hounsfield unit (HU) of the scanner must be related to the
actual electron density. A CT phantom, Catphan® 600 was
used. The Catphan® 600 has a diameter of 15 cm (longitudinal
dimension is 16 cm) and contains seven tissue substitute
materials, namely air, poly-methyl pentene (PMP), low-density
polyethylene (LDPE), polystyrene, acrylic, delrin, and teflon.
Their relative electron densities relative to water range from 0
to 1.867.
The calibration of a CT scanner involves acquiring CT
images, obtaining an average HU for each inserted material,
and plotting the HU values as a function of the relative electron
density. The calibration curve was obtained for CBCT in
full-fan and half-fan modes. The calibration curves of the CT,
full-fan CBCT, and half-fan CBCT images were compared.
Cross-sectional images of the phantom imaged with the CT,
full-fan CBCT, and half-fan CBCT are shown in Fig. 1. The
stability of CBCT calibration curve with time was also studied
over a period of 8 weeks by scanning the phantom weekly. The
HU – relative electron density curves obtained weekly were
then compared to assess the HU fluctuations with time.
(a) CT
(b) Full-fan CBCT
(c) Half-fan CBCT
Figure 1. Cross-sectional images of Catphan® 600 phantom obtained
using (a) CT, (b) full-fan CBCT, and (c) half-fan CBCT.
The CT, full-fan CBCT, and half-fan CBCT images of the
Catphan® 600 phantom were imported into the Eclipse
treatment planning system (Varian Medical Systems, Palo Alto,
CA). The HU distribution over homogeneous and
inhomogeneous media of the phantom was analyzed to quantify
the differences in the image quality of CT and CBCT. To
assess the effect of radial scatter, three homogeneous phantoms
with radial diameters of 10, 20, and 27 cm, respectively, were
scanned using CBCT.
2.3 CBCT-image-based planning
2.3.1 Isodose comparison: single direct field and wedge fields
For planning and evaluation purposes, a single direct
10 × 10 cm2, 6-MV photon beam isocentered at the center of
the Catphan® 600 phantom was generated for the CT, full-fan
CBCT, and half-fan CBCT images. Anisotropic analytical
algorithm (AAA)-based dose calculation was used and the
conventional CT electron density calibration curve was applied
for all the images. The resultant isodose curves and the dose
profiles at 1.5, 5, 10, and 15 cm depths were obtained and
compared. Similarly, four wedge fields (with angles of 15°,
30°, 45°, and 60°) of 6-MV photon beams at gantry angles of
0°, 90°, 180°, and 270° isocentered at the center of the phantom
were generated for the CT and CBCT images. The isodose
curves and dose profiles were obtained and compared.
2.3.2 Intensity-modulated radiation therapy (IMRT) dose
comparison
A detailed study using a Rando phantom was performed to
validate CBCT-based dose calculation. Three sites, namely the
head and neck, the thorax, and the pelvis, were used for the
study. The CT and half-fan CBCT images of the phantom were
obtained. For all sites, a hypothetical identical target and
sensitive structures were contoured on CT and CBCT image
sets.
For the pelvis CT, an IMRT plan with five 6-MV photon
fields, at gantry angles of 0°, 55°, 100°, 260°, and 305°,
respectively, was generated. A verification plan was created
based on CBCT images, keeping the MU values and fluence
maps of the original CT-based plan. The resultant dose
distributions and dose-volume histograms (DVHs) of the CTand CBCT-based plans were compared.
Similarly, for the head and neck CT, an IMRT plan with
seven 6-MV photon fields, at gantry angles of 0°, 50°, 100°,
150°, 310°, 260°, and 210°, respectively, was generated. For
the thorax CT, an IMRT plan with four 6-MV photon fields, at
gantry angles of 0°, 110°, 180°, and 250°, respectively, was
generated. Their corresponding verification plans were created
with the CBCT images. The resultant dose distributions and
DVHs of the corresponding plans were compared.
2.3.3 Build-up region dose comparison
In order to verify the build-up region dose distribution, at
the thoracic site, a superficial target was contoured. A
three-dimensional conformal plan using three 6-MV photon
fields, at gantry angles of 0°, 60°, and 360°, respectively,
(lateral fields with a 15° wedge) was generated. A similar plan
was created with the CBCT images. The dose distributions of
CT- and CBCT-image-based plans for the superficial target
were compared.
CBCT-image-based Dose Calculation
431
3. Results
3.1 Calibration of CT and CBCT
The HU versus relative electron density calibration curves
of the conventional CT, full-fan CBCT, and half-fan CBCT are
shown in Fig. 2. In the full-fan mode, the maximum difference
in the HU value compared with the reference CT is less than
20 HU. For the half-fan mode, HU values of tissue substitute
materials were slightly higher than those of CT and teflon
showed the maximum HU difference (40 HU). This study
confirms that CBCT has the ability to generate images with HU
values comparable to those of CT.
Figure 4. HU profiles along homogeneous and inhomogeneous media
from CT, full-fan CBCT, and half-fan CBCT images of
Catphan® 600 phantom.
Figure 2. Calibration curves (CT number versus relative electron density)
for CT, full-fan CBCT, and half-fan CBCT.
Figure 3 shows the calibration curves obtained in eight
consecutive weeks for the full-fan CBCT. No significant
variation was observed in the calibration curve, confirming the
stability of kV-CBCT with time for HU value integrity.
Figure 5 shows the HU profiles along a central axis of
three homogeneous water equivalent phantoms of various sizes.
As expected, the fluctuation range of HU values increases with
the phantom diameter, indicating the increased influence of
scatter radiation.
Figure 5. HU profiles for homogeneous water equivalent phantoms with
diameters of 27, 20, and 10 cm, respectively.
Figure 3. Stability of calibration curve over a period of 8 weeks for
full-fan CBCT.
3.2 CBCT-based planning
Figure 4 shows the HU profiles in homogeneous and
inhomogeneous media obtained from CT, full-fan CBCT, and
half-fan CBCT images of the Catphan® 600 phantom. It is
found that the HU profiles of the CT and full-fan CBCT
normally agree to within difference of 10 HU except the
peripheral region where it is reduced by 10 to 40 HU. The HU
profiles of half-fan CBCT were 10 to 30 HU higher along the
central region and 30 to 80 HU lower at the periphery when
compared to the CT profiles.
Figure 6(a) shows the dose distributions calculated using a
single direct 10 × 10 cm2 6-MV photon beam in the transverse
section of three respective sets of images. Figure 6(b) shows a
comparison of dose profiles at depths of 1.5, 5, 10, and 15 cm.
From this analysis, the dose calculated using CT agrees with
the CBCT-based dose calculation to within 1%. Figure 7(a)
shows the dose distribution calculated using wedge fields in the
transverse section of the three sets of images. Figure 7(b)
shows a comparison of dose profiles along the orthogonal beam
central axis. The CT-image-based dose calculation agrees with
that of CBCT to within 1%.
432
J. Med. Biol. Eng., Vol. 31 No. 6 2011
(i)
(ii)
(iii)
(a)
(b)
Figure 6. (a) Dose distributions calculated using a single direct 10 × 10 cm2 6-MV photon beam in the transverse section of three respective sets
of images of (i) CT, (ii) full-fan CBCT (FFCBCT), and (iii) half-fan CBCT (HFCBCT). (b) Comparison of dose profiles at depths of
1.5, 5, 10, and 15 cm.
(i)
(ii)
(iii)
(a)
(b)
Figure 7. (a) Dose distributions calculated using wedge fields in the transverse section of three sets of images of (i) CT, (ii) full-fan CBCT
(FFCBCT), and (iii) half-fan CBCT (HFCBCT). (b) Comparison of dose profiles along the orthogonal beam central axis.
CBCT-image-based Dose Calculation
CT-based plan
433
CBCT-based plan
(a)
(b)
(c)
Figure 8. (a) IMRT dose distributions computed based on CT and CBCT pelvic images. (b) Comparison of DVHs of the target and sensitive
structures. (c) Percentage dose difference computed based on Verisoft 3.1 software. Solid and dotted curves represent the results from
CT- and CBCT-based calculations, respectively.
CT-based plan
CBCT-based plan
(a)
(b)
(c)
Figure 9. (a) IMRT dose distributions computed based on CT and CBCT head and neck images. (b) Comparison of DVHs of the target and
sensitive structures. (c) Percentage dose difference computed based on Verisoft 3.1 software. Solid and dotted curves represent the
results from CT- and CBCT-based calculations, respectively.
J. Med. Biol. Eng., Vol. 31 No. 6 2011
434
CT-based plan
CBCT-based plan
(a)
(b)
(c)
Figure 10. (a) IMRT dose distributions computed based on CT and CBCT thoracic images. (b) Comparison of DVHs of the target and sensitive
structures. (c) Percentage dose difference computed based on Verisoft 3.1 software. Solid and dotted curves represent the results from
CT- and CBCT-based calculations, respectively.
CT-based plan
CBCT-based plan
(a)
(b)
Figure 11. (a) Dose distribution computed based on CT and CBCT images for superficial target. (b) Solid and dotted curves represent the results
from CT- and CBCT-based calculations, respectively.
Figures 8(a) to (c) compare the IMRT dose distributions
computed based on the CT and CBCT images in the pelvic
region of the Rando phantom. The DVHs of the target and
sensitive structures are also presented. The resultant percentage
dose difference was calculated using Verisoft 3.1 software. The
dose calculated using CT agrees with the CBCT-based dose
calculation to within ± 1%.
Figures 9 and 10 compare IMRT dose distributions
computed based on CT and CBCT images in the head and neck
site and the thorax site of the Rando phantom, respectively,
along with the respective DVHs of the target and sensitive
structures. For the head and neck site, the percentage dose
difference was found to be within ± 1%. For the thorax site, the
percentage dose difference was found to be within ± 3%, which
is clinically significant. For both cases, the doses showed good
agreement between 30% and 90% isodose lines, but there were
discrepancies at the 100% isodose line, with that for the thorax
IMRT case more pronounced.
A dose comparison of CT- and CBCT-based planning of a
superficial target is shown in Fig. 11. The dose calculated using
CT agrees with the CBCT-based calculation to within ± 1%.
CBCT-image-based Dose Calculation
4. Discussion
CBCT based on flat-panel technology integrated in a
medical linear accelerator has improved the precision of
targeting in radiotherapy. Two important applications of CBCT
are patient setup and dose verification. A CBCT image of the
patient on the treatment table can be acquired in about
60 seconds, just before delivery of each treatment. These
pretreatment volumetric images may be used to verify or correct
the patient setup and recalculate the treatment plan based on
patient anatomy on treatment day. However, this application
depends on the CBCT image quality. The quality of CBCT
images is inferior to that of conventional CT images due to
increased scatter, beam hardening, and intra-scanning organ
motion. This raises concerns about the direct use of CBCT for
dose calculation [8]. This study investigated the accuracy of
CBCT-based treatment planning using a Rando phantom for
head and neck, thorax, and pelvis sites.
Results show that the difference in HU values between CT
and CBCT images of Catphan® 600 phantom are less than
40 HU. Teflon shows the maximum variation in the half-fan
mode. There was no significant change in the HU – electron
density curve over a period of eight weeks. Thus, this study
confirms that CBCT has the ability to generate images with HU
values comparable to those of CT.
The profiles of full-fan CBCT in a homogeneous medium
were 10 to 40 HU lower at the periphery region when
compared to those of CT. For half-fan CBCT, the HU profiles
were 10 to 30 HU higher and 30 to 80 HU lower along the
central and peripheral regions, respectively, when compared to
profiles of CT. This reduction in HU at the periphery may be
due to larger primary-to-scatter ratio with a larger FOV.
The influence of HU variation in dose calculation was
evaluated by comparing the dose distribution and DVHs
between CT- and CBCT-image-based treatment planning. The
dose distributions calculated for a single direct 10 × 10 cm2
6-MV photon beam for CBCT images of Catphan®
600 phantom agrees with that of CT to within 1%. Similarly,
the dose distribution calculated with wedge fields for CBCT
images agrees with that of CT to within 1%.
In general, IMRT plans based on CBCT images for the
head and neck and pelvis sites of the Rando phantom show a
good agreement with the plans based on CT. However, for the
thorax site, a discrepancy at the 100% isodose line was noted
and the percentage dose difference was found to be
within ± 3%. For the 3D-CRT plan for a superficial target, the
percentage dose difference was found to be within 1%.
435
Alhough the results of CBCT-based plans provided in this
study are comparable to those of CT-based plans, for larger
treatment volumes, CBCT-based planning could be limited due
to the limited field size of the detector (about 16 cm in length).
However, CBCT-based planning can be useful for re-planning
purposes to provide monitor units and dose distribution or to
verify treatment delivery.
5. Conclusions
Cone-beam CT plays an important role in modern
radiotherapy in terms of accurate patient setup and dose
verification. This study investigated the dosimetric feasibility
of CBCT-based treatment planning using a Rando phantom.
Although CBCT images include larger scatter and artifacts than
do CT images, the dosimetric results of CBCT-based plans are
comparable to those of CT-based plans. Dosimetric data in
inhomogeneous tissue regions such as the thorax region should
be carefully validated.
References
[1]
[2]
[3]
[4]
[5]
[6]
[7]
[8]
L. Xing, B. Thorndyke, E. Schreibmann, Y. Yang, T. F. Li, G. Y.
Kim, G. Luxton and A. Koong, “Overview of image-guided
radiation therapy,” Med. Dosim., 31: 91-112, 2006.
D. A. Jaffray and J. H. Siewerdsen, “Cone-beam computed
tomography with a flat-panel imager: initial performance
characterization,” Med. Phys., 27: 1311-1323, 2000.
D. A. Jaffray, J. H. Siewerdsen, J. W. Wong and A. A. Martinez,
“Flat panel cone-beam computed tomography for image-guided
radiation therapy,” Int. J. Radiat. Oncol. Biol. Phys., 53:
1337-1349, 2002.
D. Yan, D. Lockman, D. Brabbins and A. Martinez, “An off-line
strategy for constructing a patient-specific planning target
volume in adaptive treatment process for prostate cancer,” Int. J.
Radiat. Oncol. Biol. Phys., 48: 289-302, 2000.
D. Letourneau, R. Wong, D. Moseley, M. B. Sharpe, S. Ansell,
M. Gospodarowicz and D. A. Jaffray, “Online planning and
delivery technique for radiotherapy of spine metastases using
cone-beam CT: image quality and system performance,” Int. J.
Radiat. Oncol. Biol. Phys., 67: 1229-1237, 2007.
R. Boopathy, S. Padmanaban, V. Nagarajan, P. Sukumaran, P.
Jeevanandam, S. Kumar, D. Rajasekaran and L. A.
Venkataraman, “Effects of lung tumor motion on delivered dose
distribution during RapidArc treatment technique,” J. Med. Biol.
Eng., 30: 189-192, 2010.
S. Padmanaban, R. Boopathy, B. Kunjithapatham, P. Sukumar
and V. Nagarajan, “A phantom study on the effects of target
motion in non-gated kV-CBCT imaging,” Australas. Phys. Eng.
Sci. Med., 33: 59-64, 2010.
Y. Yang, E. Schreibmann, T. Li, C. Wang and L. Xing,
“Evaluation of on-board kV cone-beam CT (CBCT)-based dose
calculation,” Phys. Med. Biol., 52: 685-705, 2007.
436
J. Med. Biol. Eng., Vol. 31 No. 6 2011